Streaming G-Buffer Compression for Multi-Sample Anti-Aliasing
Ethan Kerzner1;2, Marco Salvi3
1SCI Institute, 2School of Computing, 3Intel Corporation
We present a novel lossy compression algorithm for G-buffers that enables deferred shading applications with high visibility sampling rates. Our streaming compression method operates in a single geometry rendering pass with a fixed, but scalable, amount of per pixel memory. We demonstrate reduced memory requirements and improved performance, with minimal impact on image quality.
Read the preprint paper: Streaming G-Buffer Compression for Multi-Sample Anti-Aliasing [PDF 47.6MB]
Citation: Ethan Kerzner, Marco Salvi, Streaming G-Buffer Compression for Multi-Sample Anti-Aliasing, Preprint